Charles Stern
Charles Stern
According to [this comment](https://github.com/pydata/xarray/issues/5172#issuecomment-823298453), the underlying issue discussed above should be resolved with `h5netcdf >= 0.11.0`. Here's a notebook I ran today for `dataset = 'IPSL-CM6A-LR.abrupt-4xCO2.r1i1p1f1.Lmon.cLeaf.gr'` with `h5netcdf == 0.11.0`:...
> 1. _Create a test version of the recipe (see #97) and examine the total size of the test target._ Scale up based on the "pruning factor" (what fraction of...
> One way around this would be to try with copy_input_to_local_file=True. That would bypass h5py / fsspec and open the file directly from disk. Good to know. I hadn't been...
> Feel free to try with intake/filesystem_spec#694 This didn't solve it. Hit a hang with the same traceback as above on input 245 of 2117. > Also, you can call...
This issue is being reported consistently and threatens to be a major stumbling point for a smooth public launch. It is still blocking https://github.com/pangeo-forge/staged-recipes/pull/68#issuecomment-897866695. And today, in manually re-running portions...
I _think_ I have reproducer for this using xarray (rather than h5py directly) and modeled after [the `store_chunk` loop](https://github.com/pangeo-forge/pangeo-forge-recipes/blob/c1a6c475ac58269d7ac2218d77599ceb29286823/pangeo_forge_recipes/recipes/xarray_zarr.py#L581-L609). I've followed [this SO post](https://stackoverflow.com/a/494273/14658879) to setup a timeout exception to...
> were you using distributed or the simple threaded scheduler? It seems to be the latter? In that case, as a variation on the garbage collection idea, could you try...
> Perhaps we can use this as a basis for a more minimal reproducer using only gcsfs, h5py, and numpy. Here we go! Seemingly the same issue, but without xarray:...
> Before trying more library combinations, I recommend simplifying the reproducer further as outlined in #177 (comment). ...now questioning if my use of `signal` in https://github.com/pangeo-forge/pangeo-forge-recipes/issues/177#issuecomment-924337789 may be incorrectly setting...
> I think we need to get to the bottom fo the issue with a minimal reproducer. Sounds good. When I left off with it last week, I was beginning...